From: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
To: Mel Gorman <mel@skynet.ie>
Cc: linux-mm@kvack.org, ak@suse.de,
Nishanth Aravamudan <nacc@us.ibm.com>,
pj@sgi.com, kxr@sgi.com, Christoph Lameter <clameter@sgi.com>,
akpm@linux-foundation.org,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Subject: Re: [PATCH/RFC] 2.6.23-rc1-mm1: MPOL_PREFERRED fixups for preferred_node < 0
Date: Tue, 31 Jul 2007 11:58:07 -0400 [thread overview]
Message-ID: <1185897487.6240.29.camel@localhost> (raw)
In-Reply-To: <20070731153227.GB18506@skynet.ie>
On Tue, 2007-07-31 at 16:32 +0100, Mel Gorman wrote:
> On (30/07/07 18:00), Lee Schermerhorn didst pronounce:
> > On Mon, 2007-07-30 at 17:38 -0400, Lee Schermerhorn wrote:
> > > These are some "issues" that I came across working on the Memoryless
> > > Node series. I'm using the same cc: list as that series as the issues
> > > are somewhat related.
> > >
> > > Only boot tested at this point.
> >
> > I sent the wrong patch--forgot to refresh before posting :-(. Bogus
> > code in mpol_to_str() in previous patch.
> >
> > Try this one.
> >
> > Lee
> >
> > > ---------------------------
> >
> > PATCH/RFC - MPOL_PREFERRED fixups for "local allocation"
> >
> > Here are a couple of potential "fixups" for MPOL_PREFERRED behavior
> > when v.preferred_node < 0 -- i.e., "local allocation":
> >
> > 1) [do_]get_mempolicy() calls the misnamed get_zonemask() to fetch the
> > nodemask associated with a policy. Currently, get_zonemask() returns
> > the set of nodes with memory, when the policy 'mode' is 'PREFERRED,
>
> Consider a cleanup that renames get_zonemask because the naming is
> misleading.
I can do that. Wanted to hear from others, such as yourself first.
>
> > and the preferred_node is < 0. Return the set of allowed nodes
> > instead. This will already have been masked to include only nodes
> > with memory.
> >
> > 2) When a task is moved into a [new] cpuset, mpol_rebind_policy() is
> > called to adjust any task and vma policy nodes to be valid in the
> > new cpuset. However, when the policy is MPOL_PREFERRED, and the
> > preferred_node is <0, no rebind is necessary. The "local allocation"
> > indication is valid in any cpuset.
> >
> > 3) mpol_to_str() produces a printable, "human readable" string from a
> > struct mempolicy. For MPOL_PREFERRED with preferred_node <0, show
> > the entire set of valid nodes. Although, technically, MPOL_PREFERRED
> > takes only a single node, preferred_node <0 is a local allocation policy,
> > with the preferred node determined by the context where the task
> > is executing. All of the allowed nodes are possible, as the task
> > migrates amoung the nodes in the cpuset.
> >
> > Comments?
> >
> > Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
> >
> > mm/mempolicy.c | 31 ++++++++++++++++++++++++-------
> > 1 file changed, 24 insertions(+), 7 deletions(-)
> >
> > Index: Linux/mm/mempolicy.c
> > ===================================================================
> > --- Linux.orig/mm/mempolicy.c 2007-07-30 17:32:06.000000000 -0400
> > +++ Linux/mm/mempolicy.c 2007-07-30 17:38:17.000000000 -0400
> > @@ -494,9 +494,11 @@ static void get_zonemask(struct mempolic
> > *nodes = p->v.nodes;
> > break;
> > case MPOL_PREFERRED:
> > - /* or use current node instead of memory_map? */
> > + /*
> > + * for "local policy", return allowed memories
> > + */
> > if (p->v.preferred_node < 0)
> > - *nodes = node_states[N_MEMORY];
> > + *nodes = cpuset_current_mems_allowed;
> > else
>
> Is this actually a bugfix? From this context, it looks like memory
> policies using MPOL_PREFERRED can ignore cpusets.
Not a serious bug, if it is one. More of a cleanup. All this does is
return a node mask in the case where the application has a task memory
policy of 'PREFERRED with a node id of -1 [which happens when you
specify an empty nodemask to set_mempolicy() or mbind()]. This means
"local allocation"--the actual "current node id" is fetched at
allocation time. This is a little know "feature" of get_mempolicy().
The results is misleading, but there isn't much the application can do
with it. Node masks are ANDed with cpuset_current_mems_allowed when
installed via a syscall.
>
> > node_set(p->v.preferred_node, *nodes);
> > break;
> > @@ -1650,6 +1652,7 @@ void mpol_rebind_policy(struct mempolicy
> > {
> > nodemask_t *mpolmask;
> > nodemask_t tmp;
> > + int nid;
> >
> > if (!pol)
> > return;
> > @@ -1668,9 +1671,15 @@ void mpol_rebind_policy(struct mempolicy
> > *mpolmask, *newmask);
> > break;
> > case MPOL_PREFERRED:
> > - pol->v.preferred_node = node_remap(pol->v.preferred_node,
Ultimately, node_remap() [the bitmap functions it calls] will return the
old value of "-1" because it's outside the valid range for a the node
bitmasks. However, it doesn't seem right to be calling node_remap()
with an invalid node id. I think it's clearer this way:
> > + /*
> > + * no need to remap "local policy"
> > + */
> > + nid = pol->v.preferred_node;
> > + if (nid >= 0) {
> > + pol->v.preferred_node = node_remap(nid,
> > *mpolmask, *newmask);
> > - *mpolmask = *newmask;
> > + *mpolmask = *newmask;
> > + }
> > break;
> > case MPOL_BIND: {
> > nodemask_t nodes;
> > @@ -1745,7 +1754,7 @@ static const char * const policy_types[]
> > static inline int mpol_to_str(char *buffer, int maxlen, struct mempolicy *pol)
> > {
> > char *p = buffer;
> > - int l;
> > + int nid, l;
> > nodemask_t nodes;
> > int mode = pol ? pol->policy : MPOL_DEFAULT;
> >
> > @@ -1755,8 +1764,16 @@ static inline int mpol_to_str(char *buff
> > break;
> >
> > case MPOL_PREFERRED:
> > - nodes_clear(nodes);
> > - node_set(pol->v.preferred_node, nodes);
Here, I think set_bit() will set bit 31. Again, misleading, IMO.
> > + nid = pol->v.preferred_node;
> > + /*
> > + * local interleave, show all valid nodes
> > + */
> > + if (nid < 0 )
> > + nodes = cpuset_current_mems_allowed;
> > + else {
> > + nodes_clear(nodes);
> > + node_set(nid, nodes);
> > + }
> > break;
> >
> > case MPOL_BIND:
> >
>
> --
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2007-07-31 15:58 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-07-27 19:43 [PATCH 00/14] NUMA: Memoryless node support V4 Lee Schermerhorn
2007-07-27 19:43 ` [PATCH 01/14] NUMA: Generic management of nodemasks for various purposes Lee Schermerhorn
2007-07-30 21:38 ` [PATCH/RFC] 2.6.23-rc1-mm1: MPOL_PREFERRED fixups for preferred_node < 0 Lee Schermerhorn
2007-07-30 22:00 ` Lee Schermerhorn
2007-07-31 15:32 ` Mel Gorman
2007-07-31 15:58 ` Lee Schermerhorn [this message]
2007-07-31 21:05 ` [PATCH/RFC] 2.6.23-rc1-mm1: MPOL_PREFERRED fixups for preferred_node < 0 - v2 Lee Schermerhorn
2007-08-01 2:22 ` [PATCH 01/14] NUMA: Generic management of nodemasks for various purposes Andrew Morton
2007-08-01 2:52 ` Christoph Lameter
2007-08-01 3:05 ` Andrew Morton
2007-08-01 3:14 ` Christoph Lameter
2007-08-01 3:32 ` Andrew Morton
2007-08-01 3:37 ` Christoph Lameter
[not found] ` <Pine.LNX.4.64.0707312151400.2894@schroedinger.engr.sgi.com>
2007-08-01 5:07 ` Andrew Morton
2007-08-01 5:11 ` Andrew Morton
2007-08-01 5:22 ` Christoph Lameter
2007-08-01 10:24 ` Mel Gorman
2007-08-02 16:23 ` Mel Gorman
2007-08-02 20:00 ` Christoph Lameter
2007-08-01 5:36 ` Paul Mundt
2007-08-01 9:19 ` Andi Kleen
2007-08-01 14:03 ` Lee Schermerhorn
2007-08-01 17:41 ` Christoph Lameter
2007-08-01 17:54 ` Lee Schermerhorn
2007-08-02 20:05 ` [PATCH/RFC/WIP] cpuset-independent interleave policy Lee Schermerhorn
2007-08-02 20:34 ` Christoph Lameter
2007-08-02 21:04 ` Lee Schermerhorn
2007-08-03 0:31 ` Christoph Lameter
2007-08-02 20:19 ` Audit of "all uses of node_online()" Lee Schermerhorn
2007-08-02 20:26 ` Christoph Lameter
2007-08-08 22:19 ` Lee Schermerhorn
2007-08-08 23:40 ` Christoph Lameter
2007-08-16 14:17 ` [PATCH/RFC] memoryless nodes - fixup uses of node_online_map in generic code Lee Schermerhorn
2007-08-16 18:33 ` Christoph Lameter
2007-08-16 19:15 ` Lee Schermerhorn
2007-08-16 21:10 ` Lee Schermerhorn
2007-08-16 21:13 ` Christoph Lameter
2007-08-24 16:09 ` [PATCH] 2.6.23-rc3-mm1 - Move setup of N_CPU node state mask Lee Schermerhorn
2007-09-06 13:56 ` Mel Gorman
2007-08-02 20:33 ` Audit of "all uses of node_online()" Andrew Morton
2007-08-02 20:45 ` Lee Schermerhorn
2007-08-01 15:58 ` [PATCH 01/14] NUMA: Generic management of nodemasks for various purposes Nishanth Aravamudan
2007-08-01 16:09 ` Nishanth Aravamudan
2007-08-01 17:47 ` Christoph Lameter
2007-08-01 15:25 ` Nishanth Aravamudan
2007-07-27 19:43 ` [PATCH 02/14] Memoryless nodes: introduce mask of nodes with memory Lee Schermerhorn
2007-07-27 19:43 ` [PATCH 03/14] Memoryless Nodes: Fix interleave behavior Lee Schermerhorn
2007-07-27 19:43 ` [PATCH 04/14] OOM: use the N_MEMORY map instead of constructing one on the fly Lee Schermerhorn
2007-07-27 19:43 ` [PATCH 05/14] Memoryless Nodes: No need for kswapd Lee Schermerhorn
2007-07-27 19:43 ` [PATCH 06/14] Memoryless Node: Slab support Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 07/14] Memoryless nodes: SLUB support Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 08/14] Uncached allocator: Handle memoryless nodes Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 09/14] Memoryless node: Allow profiling data to fall back to other nodes Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 10/14] Memoryless nodes: Update memory policy and page migration Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 11/14] Add N_CPU node state Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 12/14] Memoryless nodes: Fix GFP_THISNODE behavior Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 13/14] Memoryless Nodes: use "node_memory_map" for cpusets Lee Schermerhorn
2007-07-27 19:44 ` [PATCH 14/14] Memoryless nodes: drop one memoryless node boot warning Lee Schermerhorn
2007-07-27 20:59 ` [PATCH 00/14] NUMA: Memoryless node support V4 Nishanth Aravamudan
2007-07-30 13:48 ` Lee Schermerhorn
2007-07-29 12:35 ` Paul Jackson
2007-07-30 16:07 ` Lee Schermerhorn
2007-07-30 18:56 ` Paul Jackson
2007-07-30 21:19 ` Nishanth Aravamudan
2007-07-30 22:06 ` Christoph Lameter
2007-07-30 22:35 ` Andi Kleen
2007-07-30 22:36 ` Christoph Lameter
2007-07-31 23:18 ` Nishanth Aravamudan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1185897487.6240.29.camel@localhost \
--to=lee.schermerhorn@hp.com \
--cc=ak@suse.de \
--cc=akpm@linux-foundation.org \
--cc=clameter@sgi.com \
--cc=kamezawa.hiroyu@jp.fujitsu.com \
--cc=kxr@sgi.com \
--cc=linux-mm@kvack.org \
--cc=mel@skynet.ie \
--cc=nacc@us.ibm.com \
--cc=pj@sgi.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox